Fully automated AI-based quantification of tumour volume on PSMA PET-CT images is associated with overall survival in prostate cancer.
Authors
Affiliations (7)
Affiliations (7)
- Department of Translational Medicine, Lund University, Malmö, Sweden.
- Department of Clinical Physiology and Nuclear Medicine, Skåne University Hospital, Malmö, Sweden.
- Department of Urology, Skåne University Hospital, Malmö, Sweden.
- Chalmers University of Technology, Gothenburg, Sweden.
- Eigenvision AB, Lund, Sweden.
- Department of Radiation Physics, Skåne University Hospital, Malmö, Sweden.
- Department of Molecular and Clinical Medicine, Gothenburg University, Gothenburg, Sweden.
Abstract
Artificial intelligence (AI)-based models can quantify tumour volume in positron emission tomography (PET) examinations using prostate specific membrane antigen (PSMA). It is, however, unknown whether this measurement of tumour volume contains prognostic information. The purpose of this study was to examine whether tumour volume measured by AI is associated with overall survival in patients with prostate cancer. Patients who had undergone [<sup>18</sup>F]PSMA-1007 PET-CT due to primary staging or biochemical recurrence (BCR) of prostate cancer at our centre were retrospectively included and followed for 5 years. Automatic quantification of tumour volume was performed using a previously developed fully automated AI model from our research group for PSMA PET-CT analysis, without manual corrections. The tumour volume for the prostate, lymph nodes, and bone was calculated and summarized as Total Lesion Volume (TLV) and divided categorically (no detectable tumour, low, and high) as well as aggregated to whole-body TLV and divided into tertiles (low, intermediate, and high). Mortality rates were compared using Cox-regression and adjusted for age. Among 267 patients in the final analysis, 48 (18%) died during follow-up. Whole-body TLV was significantly associated with mortality in the combined cohort (p = 0.001) and in the primary staging cohort (p = 0.014) but did not reach statistical significance in the BCR cohort (p = 0.061). Fully automated AI-derived TLV from PSMA PET-CT was associated with overall survival. These findings support the potential role of AI-based tumour quantification as a potential prognostic biomarker in prostate cancer.